Hands-On Automated Machine Learning

Bagging

Bagging is also known as bootstrap aggregation. It is a way to decrease the variance error of a model's result. Sometimes the weak learning algorithms are very sensitive—a slightly different input leads to very offbeat outputs. Random forest reduces this variability by running multiple instances, which leads to lower variance. In this method, random samples are prepared from training datasets using the random sample with replacement models (bootstrapping process).

Models are developed on each sample using supervised learning methods. Lastly, the results are merged by averaging the predictions or selecting the best prediction utilizing the majority voting technique. Majority voting is a process in which the prediction of the ensemble ...